activities at the interfaces with chemistry, engineering and biology. With a total of 450 staff including more than 100 PhD students, it is one of the

activities at the interfaces with chemistry, engineering and biology. With a total of 450 staff including more than 100 PhD students, it is one of the

largest physics laboratories in France. Our major research fields are:

• Magnetism and spintronics,

• Photonics, plasmonics and non-linear optics,

• Correlated systems, quantum fluids and superconductivity,

• Quantum, molecular and wide bandgap electronics,

• Materials: their synthesis, structure and functions,

• Instrumentation.

These interlinked domains are developed across our three scientific departments, which are constituted from a total of 17 research teams supported by a similar number of transverse technology teams and services.

This 2015 issue of the Institute’s “ Highlights ” magazine contains sixteen short articles specially written to describe particularly important scientific or technical advances made recently in our laboratory. This is of course far from being an exhaustive list, and a much more complete overview of our activities can be found at our website www.neel.cnrs.fr, together with pdf versions of all previous issues of the “ Highlights ” magazine.

The present issue also features an article describing our visitors’ centre, the “PhysIQuarIuM”, which was created for an audience wider than this magazine’s readership. It offers a pedagogical presentation of our laboratory’s research subjects and provides the opportunity for visitors to perform simple experiments. Opened in 2014, it now attracts a large number of high-school classes and other visitors.

The year 2016 will bring considerable changes for the Institut NÉEL, at two levels. at the internal level, a new team of directors will take over from the team formerly led by alain schuhl and at present by myself. at another level, three of Grenoble’s universities will be amalgamated in 2016 to create the pluridisciplinary université Grenoble alpes, and a “Communauté université Grenoble alpes” will coordinate university, engineering school and public laboratory research across the Grenoble and alpine area. Thus, new collective challenges of great importance are ahead of us, which will make this new year (and following years) additionally stimulating for Institut NÉEL staff.

I wish you interesting reading in this 2015 issue of our Highlights magazine.

Editorial

Hervé COURTOIS

Director of Institut NÉEL

Grenoble, France

Measuring the phase shift of an electron scattered by a single spin 4 Imaging interfaces in artificial semiconductor crystals 5

TeraHertz properties of multiferroic compounds 6

Implementing X-Ray diffraction/scattering tomography in the lab 7 Strong round ceramic superconducting wires for very high field magnets 8 Metal colloids investigated by enhanced-Raman correlation spectroscopy 9 The PHySIquaRIuM, a visitors’ area at the Institut NÉEL 10 Probing light fields with nano-mechanical oscillators 11

Chiral domain walls in thin magnetic films 12

Imaging inside Gallium Nitride wires for next generation blue LEDs 13 Coherent control of a single nuclear spin with an electric field 14 Resistivity, magnetism and shape memory in martensitic alloys 15 Synchrotron X-Ray study of platinum-nanoparticle catalysts 16 Multifunctional substrate holder for a film-deposition system 17 Creation and manipulation of magnetic monopoles in spin ice 18 Long-distance entanglement between spins in quantum spin chains 19 2D nanoporous graphene: a route to organic topological insulators? 20

DIRECTOR OF PUBLICATION Hervé COURTOIS

EDITORS Signe SEIDELIN,

Jan VOGEL, Ronald COX

Contents

PRODUCTION mANAGER

Nathalie BOURGEAT-LAmI

LAYOUT Élodie BERNARD Florence FERNANDEZ

PRINTING Press’Vercors November 2015

Imaging interfaces in artificial semiconductor

crystals

Quantum interference is one of the most mysterious features of quantum mechanics. A well-known example is the Young double-slit experiment which demonstrates that light can be described at the same time by a particle picture and by a wave picture – the famous particle-wave duality. The same can be said of electrons when they pass through a two-path interferometer. In this article, we show that an electron interferometer can be used to measure a very intriguing effect postulated to arise when electrons are scattered off a magnetic impurity at temperatures near the zero temperature limit. In this case the electrons should acquire a phase shift of p/2, contrary to what is observed when an electron scatters off a static impurity. This phase shift, predicted as early as 1974 has finally been measured using a micron scale, quantum interferometer.

measuring the phase shift of an electron

scattered by a single spin

physicist Philippe Nozières in 1974. This predicted phase shift, however, has stayed elusive for 40 years. This is because one has to be able to isolate a single magnetic impurity to demonstrate the effect. That requires sophisticated nanofabrication techniques that have only become available in the last decade. In addition, in order to detect the phase shift, phase sensitive measurements similar to the Young double slit experiment have to be realized for a system of electrons.

A single localized electron can scatter mobile electrons exactly like a single magnetic impurity, and it can be controlled more easily than an impurity atom. A single electron can now be trapped routinely within a sub- micron electrostatic trap, called a “quantum dot”, defined in a two-dimensional electron gas by carefully shaped electrodes. In this device, a Kondo cloud will form in the surrounding electron gas.

By inserting such a quantum dot in one arm of a quantum interferometer (Fig. 2), we have been able to measure the predicted p/2 phase shift (Fig. 1) in the regime of the Kondo scattering, which sets in at temperatures of the order of one hundredth of a degree Kelvin for this system.

Using powerful computing algorithms to compare our measurements with exact numerical results of modern theory, we have shown that this scenario, based on the Kondo effect, is correct. The understanding of electrical transport in metals has thus reached a new level: the most modern theoretical models appear to be essential for a complete comprehension of phenomena as basic as the electronic transport governing the resistance of metals.

The electrical resistance of a pure metal usually drops when the temperature is lowered because electrons can travel through the crystal more easily when lattice vibrations are reduced. However below a temperature of approximately 10 K the resistance stops falling, stabilising at a level determined by scattering by lattice defects in the material. Upon further cooling, some metals – for example lead or aluminium – can suddenly lose their

electrical resistance and become super-conducting, but

“normal” metals, like copper, silver or gold do not show superconductivity and have a finite resistance even at the lowest accessible temperatures.

When normal metals contain a tiny amount of magnetic impurities, such as iron or cobalt, they show a different, very unusual behaviour. In this case the electrical resistance does not stay constant when the temperature approaches absolute zero, but it increases. This effect puzzled physicists over many decades and it became known as the Kondo effect, after the Japanese physicist Jun Kondo who was the first to give an explanation of the unusual resistance increase in 1964.

Electrons carry not only an electric charge but also a magnetic moment associated with their spin. The interaction of this magnetic moment with magnetic impurities generates collisions of a novel type, which become more and more efficient as the temperature is lowered. The result of these interactions is the formation of a large static cloud of electrons, often called the Kondo cloud. At very low temperatures (well below 1 deg. K), this cloud of electron spins screens the magnetic impurity and acts like a large, static “impurity”, scattering other electrons passing nearby. This is the Kondo effect.

It is believed that, when an electron is scattered by the Kondo cloud, the electron’s wave function acquires a phase shift of p/2 (see Fig.1) as predicted by the French

Fig. 2: Quantum interferometer made by depositing electrostatic gates onto a two-dimensional electron gas in a semiconductor device. The blue dotted lines depict the two possible paths for electrons passing through the interferometer. A quantum dot containing a single localised electronic spin (in yellow), which acts as an effective Kondo “impurity“, is inserted in one of the arms of the interferometer.

Fig. 1: Phase shift across a Kondo

“impurity“: a foreign atom or a localised electron. Above a characteristic temperature T_K associated with the Kondo effect the incident electron wave is scattered without any phase shift. Below this temperature, a p/2 phase shift due to the Kondo effect has been observed for the first time in this work.

CONTACT

Christopher BÄUERLE

christopher.bauerle@neel.cnrs.fr Tristan MEUNIER

tristan.meunier@neel.cnrs.fr Ph.D student:

Shintaro TaKaDa

FURTHER READING

“Transmission phase in the Kondo regime revealed in a true two-path interferometer”

S. Takada, C. Bäuerle, M. yamamoto, K. Watanabe, s. hermelin, T. Meunier, a. alex, a. Weichselbaum, J. von Delft, a. D. Wieck and s. Tarucha Phys. Rev. Lett. 113, 126601 (2014).

“Measurement of the Transmis- sion Phase of an Electron in a Quantum Two-Path Interfero- meter”

s. Takada, M. yamamoto, C. Bäuerle, K. Watanabe, a. Ludwig, a. D. Wieck and S. Tarucha

Appl. Phys. Lett. 107, 063101 (2015).

Superlattices built from two compounds with no atom in common exhibit special optical and electronic properties.

We have been investigating Zinc Telluride / Cadmium Selenide (ZnTe/CdSe) short-period superlattices, as potentially efficient absorbers of sunlight for solar cells.

Our superlattices are alternating thin layers of ZnTe and CdSe with repeat period of order 10 nm. They are grown by Molecular Beam Epitaxy (MBE), the technique of choice to realize semiconductor hetero-structures with flat, sharp interfaces. In this context a precise knowledge of the interfaces is crucial for “engineering” the optical absorption threshold: The short period makes the optical absorption properties very sensitive to the atomic arrangement at the interface.

To take the simplest possibilities, at the interface, do we have Zn atoms bound to Se, or Cd atoms bound to Te? Or more complex configurations? To answer these questions, we used a state-of-the-art Transmission Electron Micro- scope, located at the Nano-Characterization Platform (PFNC) at CEA-Minatec (Grenoble), whose resolution has been enhanced to better than 0.1 nm by correcting magnetic-lens aberrations. We used a mode where the image contrast is directly related to the number of electrons Z carried by each atom, thus distinguishing different atoms: the higher the number Z, the brighter the image.

High Resolution Transmission Electron Microscopy relies on viewing a very thin slice of the sample precisely parallel to columns of atoms in a chosen direction of the crystal structure. The transmission image shown in Fig. 1 is a view looking parallel to the plane of the CdSe/

ZnTe interface. In this projection of the structure, the images of the closest columns of metal and non-metal atoms are only partly resolved; they appear as a pair of unequal dots that blur into each other. But the dots can be clearly identified as two atoms with different values of their electron number Z. The respective brightness of the two dots is inverted on each side of an interface (A or B in Fig.1): In the image, one can differentiate high brightness/low brightness dots for CdSe (Z_Cd=48, Z_Se=34) and low brightness/high brightness dots for ZnTe layers (Z_Zn=30, Z_Te=52), see the insets at top of Fig. 1.

At the “ZnTe on CdSe” interface A, we see low brightness/

low brightness dots. Their quantitative intensity profile (Fig. 1) is consistent with Zn-Se bonds (Z_Se=34, Z_Zn=30).

That is, the interface is a very thin slice of a third compound ZnSe, at the atomic scale (Fig. 2). This implies that, when the incident molecular beam is permuted from Cd+Se to Zn+Te during superlattice growth, the change in non- metal species (Te replacing Se) is retarded compared to the change in the metal species (Zn replacing Cd). The

“CdSe on ZnTe” interface also appears to consist of a ZnSe-rich layer. High brightness/high brightness dot pairs never appear in the image of the superlattice, ruling out the presence of a CdTe interface layer.

many electronic and optoelectronic devices are fabricated from semiconducting “superlattices”: a sequence of alternating, thin layers of two different semiconducting compounds that have the same

crystal structure. A superlattice is like a single crystal, but one that has periodic, abrupt changes of composition at the interfaces between its two constituent compounds. Usually these are compounds of a metal and a nonmetal where only the metal component is modulated (e.g. Gallium Arsenide/

Aluminium Arsenide superlattices, a classic example). We have been studying the interesting case of superlattices where both the metal atoms (the cations) and the non-metal atoms (the anions) are different in the alternate layers. The atomic configurations at the interfaces between two layers is then intrinsically complex, but can be unravelled by ultra-high resolution electron microscopy.

Imaging interfaces in artificial semiconductor

crystals

To add additional chemical insight to the image contrast analysis, the samples were investigated by Atom Probe Tomography (APT), another technique available at the Nano-Characterization Platform. A 3-Dimensional chemical mapping of atomic species can be done with (lower) spatial resolution by evaporating, atom by atom, a sharp tip fashioned in the material. ZnSe⁺ molecular ions were detected, originating precisely from all the interfaces, but not CdTe molecular ions.

The predominant formation of ZnSe bonds at both interfaces proves to be a very robust phenomenon: It persists under a wide range of crystal growth conditions investigated in this study. It is attributed to the difference in cohesive energies between the four possible metal/

non-metal couples, which clearly favours ZnSe.

Fig. 1: High resolution electron microscopy image of several layers of a CdSe/ZnTe superlattice.

Growth direction is from left to right. Interfaces ZnTe-on-CdSe and CdSe-on-ZnTe are labelled A and B, respectively. Each bright spot represents a close pair of columns in the crystal structure (a column of metal atoms and a non-metal atom column).

The two zooms show that the blurred bright “spots“ consist of weak dot/bright dot pairs, corresponding to either (Cd, Se) or (Zn, Te) columns.

At bottom: brightness profile across the interface A showing the Zn-Se bond nature of the interface.

Fig. 2: Atomic structure deduced for the interface between Cadmium Selenide (at left) and Zinc Telluride (at right), showing a Zn-Se bond at the interface.

(Colour code: red=Zn, blue=Se, yellow=Cd, green=Te).

CONTACT

Catherine BOUGEROL catherine.bougerol@neel.cnrs.fr Ph.D student:

Bastien BONEF

FURTHER READING

“Atomic arrangement at ZnTe/CdSe interfaces determined by high resolution scanning transmission electron microscopy and atom probe tomography”

B. Bonef, L. Gérard, J-L. rouvière, a. Grenier, P-h. Jouneau, E. Bellet-amalric, h. Mariette, r. andré and C. Bougerol App. Phys. Lett. 106, 051904 (2015).

Electromagnetic waves outside the spectrum of visible light have been used for over a century to carry information (radio waves) or probe matter (X-Rays). However the spectral range known as the “TeraHertz gap” (Fig. 1) has hardly been explored because of a lack of good sources and detectors. Broadband spectroscopy at TeraHertz frequencies has now become available at synchrotron radiation sources. This spectral range is of considerable interest in Condensed matter Physics, especially in magnetic or dielectric compounds, since both magnetic excitations (called magnons) and their electric analogue (optical phonons) have energies in this range. We have explored the TeraHertz response of compounds where new phenomena are expected, the so called “electro-magnons”.

TeraHertz properties of multiferroic

compounds

magnetic, an excitation at the same energy is observed but polarization measurements show that it is excited only by the electric field, along the c-axis, not at all by the magnetic field!

One can also study THz excitations with neutron scattering. Since neutrons carry a magnetic moment, they can distinguish magnetic excitations from atomic excitations. Our neutron scattering measurements have confirmed the magnetic nature of the 1.2 THz (5.2 meV) excitation. These combined results demonstrate that a Manganese magnon has become electro-active in the Erbium compound. This is a new kind of excitation called an “electro-magnon”. We attribute its electro-activity to the presence of the second magnetic ion in the material, the rare-earth Erbium, as follows.

The manganite compounds of formula RMnO₃, where R is the transition element Y (Yttrium) or a rare earth element such as Dysprosium, Holmium, Erbium, are ferroelectric at room temperature: They show a spontaneous electric-charge ordering, giving an electric polarization directed along the principal axis (the c-axis) of the crystal’s hexagonal lattice structure. These materials also present an antiferromagnetic order, typically below 80 K, associated with the Mn³⁺ ions (see Fig. 2). Compounds of this type, where both an electric and a magnetic ordering prevail simultaneously, are called “multiferroics”. They are attracting considerable attention since they have potential applications for recording and manipulating information using both electric and magnetic fields.

We have focused on two members of this family, namely YMnO₃ and ErMnO₃. We have fully characterized their energy spectra in the TeraHertz (THz) range by combining two techniques available at large-scale facilities: THz spectroscopy at the synchrotron SOLEIL near Paris and inelastic neutron scattering at the Institut Laue-Langevin (ILL) in Grenoble. Both techniques allow measurements of the THz spectra, but bring different information.

Since the THz electromagnetic wave has perpendicular electric and magnetic field components, one can differentiate electric and magnetic excitations by their polarization dependence. That is, one can distinguish magnons, which are associated with the magnetic order and excited by the THz magnetic field, from optical phonons, associated with the ordered lattice of positive and negative ions and excited by the THz electric field.

In YMnO₃, where the Y³⁺ ion is non magnetic, a sharp excitation is observed at 1.2 THz whenever the THz magnetic field is polarized perpendicular to the c-axis, for all orientations of the perpendicular electric field. It has all the characteristics of a magnon corresponding to the precession of the Mn magnetic moment around the c-axis. In ErMnO₃, where the rare earth ion Er³⁺ is

Fig. 2: ErMnO3 magnetic and electric orders.

The electric polarisation is along the c-axis. The Mn magnetic moments are oriented at 120° from each other.

Fig. 1: Electromagnetic spectrum from radio- waves to X-Rays. In the THz range, probed at SOLEIL and ILL, both atomic vibrations (phonons) and magnetic excitations (magnons and crystal-field excitations) can be studied.

CONTACT

Sophie DE BRION sophie.debrion@neel.cnrs.fr Virginie SIMONET virginie.simonet@neel.cnrs.fr Ph.D student:

Laura CHaIX

FURTHER READING

“Magneto- to electroactive transmutation of spin waves in ErMnO₃”

L. Chaix, S. de Brion, S. Petit, r. Ballou, L.-P. regnault, J. Ollivier, J.-B. Brubach, P. roy, J. Debray, P. Lejay, a. Cano, E. ressouche and V. Simonet

Phys. Rev. Lett. 112, 137201 (2014).

It is known that the electronic levels in rare earth ions are shifted and split by the electrostatic field of the surrounding ions, the so-called “crystal field”.

Transitions between the crystal field-split levels may be allowed depending on the local symmetry. Thanks to our complementary probes, THz waves and neutrons, we have identified six such transitions for ErMnO₃. One of these transitions, at 2.1 THz, has the correct symmetry to be electro-active and to couple strongly to the magnon at 1.2 THz. The coupling interaction is then responsible for the complete loss of the magnetic character of the magnon: It has been transmuted into a purely electro-active excitation. This new mechanism for electro-magnons may be general to other rare-earth- based multiferroics and suggests new possibilities for manipulating these excitations through the action of magnetic and electric fields.

Tomography (from the Greek “tomos“ = “slice”) is a method for generating virtual images of slices through an object. In Diffraction and Scattering Computed Tomography (DSCT), the object is mounted in a very narrow beam of monochromatic X-Rays, and the diffraction and scattering intensities are recorded with a 2-Dimensional (i.e. multi- pixel) detector, see Fig 1. This is done with the object oriented at a large number N_w of rotational positions w, for each of a similarly large number N_y of horizontal positions y. From this very large mass of data, one can compute an image of the slice that contains structural information: chemical composition, atomic order (crystalline or amorphous, type of phase), grain size, etc. If desired, the object can then be translated vertically in steps to build a 3D representation as a stack of slices. DSCT also increases sensitivity to weak X-Ray signals coming from minor phases.

To implement DSCT in-laboratory, with spatial resolution

~200 microns, we adapted a multi-circle goniometer normally dedicated to single-crystal diffraction measurements and we added the translational and rotational stepping motions required for tomographic data acquisition. To acquire one slice, we record N_y x N_w (N~100) diffraction images on a multipixel 2D detector.

The various, randomly oriented, lattice-planes (hkl) in polycrystalline material diffract the incident X-Ray beam through their specific Bragg deviation angles 2Q, giving the concentric rings in Fig. 1. Any highly disordered or amorphous materials scatter diffusely.

As illustration, we discuss our analysis of a carbon sample resulting from the phase transformation of Carbon 60 (C₆₀ “fullerene” molecules) into a mixture of a crystalline rhombohedral polymer (C60R) and disordered graphite by high pressure (5 GPa) at 1100 K. Each of the N_y x N_w 2D diffraction/scattering patterns is integrated azimuthally around its centre to convert it into a 1D pattern, i.e.

diffraction/scattering intensity vs the deviation angle 2Q, like that of Fig. 2a where we have labelled strong Bragg

Real materials often exhibit a complex and heterogeneous organization. Their distinctive functionalities are linked to their specific structure at the atomic scale as observed in alloys, cements but also in teeth, bones or archaeological materials. We have developed a method combining the diffraction and scattering of X-Rays with computed tomography. The diffraction/scattering information is used to detect, identify and image the various crystalline/amorphous phases in heterogeneous samples. This non-destructive analysis technique, first developed on synchrotron radiation facilities, can now be implemented at medium resolution on ordinary laboratory X-Ray diffraction equipment.

Implementing X-Ray

diffraction/ scattering tomography

in the lab

diffraction peaks for C60R and disordered graphite.

From the complete set of N_y x N_w1D diffraction patterns, we can compute virtual slices that image the intensity scattered at a specific Bragg angle, versus x and y at height z in the sample. Additionally a reverse analysis of the transformed data can be performed in order to extract pure single-phase diffraction patterns (e.g. Fig.

2b) for all the phases within the heterogeneous sample, including unexpected components.

For our carbon sample (Fig. 3), we evidenced (i) rhombohedral C60 domains in the middle and lower part of the sample and (ii) disordered graphite domains in the external and top part of the sample. Interestingly (iii) a hexagonal Boron Nitride (h-BN) impurity phase coming from the high pressure gasket was also localized.

This work demonstrates that X-Ray diffraction/scattering tomography can be performed in the laboratory, with all the advantages of easy and instant access that this provides, even if the spatial resolution presently attainable (~100 microns) is very much lower than that obtainable at a Synchrotron radiation facility. Our technique will be particularly interesting for samples containing crystallites and/or heterogeneities a few hundred microns in size. It is suited for numerous practical applications like cements, alloys, catalysts, bones and paint cross-sections.

CONTACT

Olivier LEyNAUD olivier.leynaud@neel.cnrs.fr Pauline MARTINETTO pauline.martinetto@neel.cnrs.fr Ph.D students:

Michelle ÁLvaREz-MuRGa Sophie CERSOy

FURTHER READING

S. Cersoy, O. Leynaud, M. Álvarez-Murga, P. Martinetto, P. Bordet, N. Boudet, E. Chalmin, G. Castets and J.-L. hodeau J. Appl. Cryst. 48, 159 (2015).

M. Álvarez-Murga, P. Bleuet and J.-L. hodeau

J. Appl. Crystallogr. 45, 1109 (2012).

Fig. 1: Geometry for X-Ray diffraction and scattering tomography analysis. To analyse one slice, the sample is translated in steps along y and rotated 180 deg. through angle w.

Fig. 3: At left: The transformed carbon sample within its glass capillary container.

At right: Three phase-selective virtual slices as computed from the data for specific Bragg angles corresponding to (i) the (003) planes of the rhombohedral C60 component, (ii) the (002) planes of disordered textured graphite, and (iii) an unanticipated hexagonal-BN impurity. (The sample shape is outlined in yellow.)

Fig. 2: (a) A global 1D diffraction pattern for a slice through the transformed carbon sample shows Bragg diffraction peaks for lattice planes of its Rhombohedral C60 and disordered textured graphite (DG) components.

(b) Reverse analysis yields 1D-diffraction patterns selective for (in green) Rhombohedral C60 and (in red) disordered graphite.

High magnetic fields have many advanced applications. In medical magnetic Resonance Imaging they make it possible to scan the soft tissues of the human body, in Nuclear magnetic Resonance spectroscopy, they enable studies of molecular structure and molecular interactions. magnetic fields curve the paths of charged particles in circular particle accelerators like the Large Hadron Collider at CERN. They serve to confine the plasma in experimental thermonuclear reactors that produce energy by the same fusion reactions as occur in the sun. These and other applications are making continuing demands for more and more powerful magnets.

Strong roundceramic superconducting wires for very high

field magnets

al., Nature Materials 2014). Thus Bi-2212 round wire is now an attractive option for high field magnets. However its mechanical properties are not yet good enough to withstand the extremely high Lorentz forces experienced when it carries a high current under a strong magnetic flux density. An additional problem is that the oxygenation heat-treatment process for Bi-2212 must often be done after winding the wire into its final shape.

In collaboration with the electrical-cable company Nexans- France, we have studied and implemented an original solution for mechanically reinforcing round Bi-2212 wires. The reinforcement is performed by a metal sheath wrapped around the wire using the Nexans company’s shaping and welding process. This intricate and delicate process consists in mechanically shaping a thin strip of metal around the wire, then welding it by laser. Finally, the sheath is stretched to adjust to the wire diameter.

The 0.8 mm diameter Nexans Bi-2212 wire employed has seven sub-elements, each with 85 filaments of superconducting material, see Figure 2a. These filaments are embedded in a matrix of pure silver with a stronger alloy MgAg for the outer part. Several metals for the sheath were studied to determine their resistance to the heat treatment of the Bi-2212 and their mechanical properties after the treatment. We chose the austenitic, nickel-chromium based “superalloy” Inconel 601. It shows a good weldability, an acceptable ductility at room temperature, and it degrades the wire’s critical current I_c only very slightly. To enable the oxygenation of the Bi-2212 during the heat-treatment, we developed a process of perforating the metal sheath in a “zig-zag”

pattern (Fig. 2b and 2c) that minimizes weakening of the metal. The 50 µm thick external sheath increases the wire’s mechanical stress resistance by a factor of 2.1 while reducing J_ce measured at 4 K by only 21%.

These results validate our reinforcement method, an essential step in making Bi-2212 wires a feasible solution for building big magnets for fields to 20 T and beyond.

Very high magnetic fields can be produced by large electric currents circulating in a copper-wire magnet, but these currents dissipate energy by the Joule effect. The losses can be enormous. For example, at the Grenoble High Magnetic Field Laboratory, 24 MW are dissipated to produce a flux density of 36 Tesla in a 34 mm dia meter bore. 24 MW is the average power consumption of a European town of 24 000 inhabitants! The solution for dramatically reducing the power consumption has been to use metallic superconductors, which have zero Joule losses below a critical temperature, T_c , near absolute zero. But the magnetic flux density is limited to a certain value, the “critical field” B_c , that destroys the super- conductivity. Even the best metallic super conductors (niobium-titanium and niobium-tin alloys) cannot satisfy future needs for higher fields.

Advances in High Temperature Superconductors are changing the situation. These ceramic materials, with values of T_c up to around 100 K, can outperform metallic superconductors when used at liquid Helium temperature. In particular, “YBCO” (Yttrium Barium Copper Oxide) and “Bismuth-2212” (Bismuth Strontium Calcium Copper Oxide, Bi₂Sr₂CaCu₂O_x) can remain superconducting up to very high current densities at 4 K, even under very high fields B, see Fig. 1.

YBCO can sustain very high currents, but has several drawbacks, notably its shape: it is supplied as very thin tapes. The Bi-2212 oxide can be produced as round wires, much better suited to building a magnet. And this compound has another big advantage: its absence of anisotropy. For YBCO, when the magnetic field direction is transverse (out of the plane of YBCO tape), the engineering critical current density J_ce decreases very rapidly (Fig. 1).

Recently the current carrying capacity of the Bi compound has been greatly enhanced using a new oxygenation process under high overpressure (D. C. Larbalestier et

Fig. 2: (a) Cross-section of a 0.8 mm diameter multi- filament Bi-2212 wire with our 50 µm thickness reinforcing sheath in Inconel (blue). (b) Inconel strip material with zigzag pattern of holes. (c) Tube-like sheath formed by rolling the Inconel strip round the wire and stretching it to fit; the holes (elongated by the stretching) enable oxidising heat treatment of the wire.

Fig. 1: Engineering critical current density J_ce at liquid helium temperatures for superconducting wires used for winding high field magnets.

(The “engineering“ critical current density is the maximum sustainable current per unit area of the total cross-section of the wires, i.e. including non-superconducting structural and reinforcing materials.)

CONTACT

Pascal TIXADOR

pascal.tixador@neel.cnrs.fr

FURTHER READING

“Mechanically reinforced Bi-2212 strand”

P. Tixador, C.E. Bruzek, B. Vincent, a. Malagoli, X. Chaud IEEE Trans. Appl. Supercond. 25, 6400404 (2015).

The rich optical properties of metal colloids are governed by the localized surface-plasmon modes of their constituent nanoparticles. These resonances are collective oscillations of the free electrons at the surface of the metal. They can be excited by light, which gives rise to resonantly-enhanced light absorption and emission properties as well as enhancement of the electric and magnetic fields at the nanoparticle surface.

Depending on various physical or chemical parameters (temperature, pH, adsorbed molecules...) nanoparticles may aggregate to form clusters ranging from simple dimers up to large and (more or less dense) fractal clusters composed of hundreds of particles. The electromagnetic interactions between the nanoparticles induce additional and stronger plasmon resonances than those of the individual particle, leading to possibly-total absorption of the incident light or to giant enhancements of the electromagnetic field in the nanoscale regions in between the particles.

These enhanced fields are in great part responsible for the so-called Surface-Enhanced Raman Scattering (SERS) effect, the enhancement by several orders of magnitude of the Raman scattering (i.e. the inelastic scattering of light) by molecules when they are adsorbed on the nanoparticles and trapped in such high field regions.

As a consequence, metal colloids are currently used for ultra-sensitive detection of molecules, with wide applications in physics, chemistry, and biology. They are also very promising prospective candidates for in-vivo medical applications, such as intracellular imaging or in-body drug delivery, and are under investigation in the field of nanoscale thermally-activated processes such as nanocatalysis or photothermal therapy.

If the case of dimers is well understood, adding more nanoparticles rapidly increases the complexity of the system, hence posing a non-trivial task to relate their measured optical properties to those of an individual nanoparticle. There is consequently a need for in-situ characterizations of the aggregates and the aggregation processes, including the very first steps of the aggregation process.

To achieve this, we have developed an original experimental technique that associates correlation analysis with SERS spectroscopy (Fig. 1). A laser beam is tightly focused in the colloidal solution via a confocal microscope thus providing a very small observation volume. The laser beam excites the nanoparticles’

plasmon-resonances responsible for enhancing the Raman scattering associated with excitations of the vibrations of the molecules adsorbed on these particles.

We record the fluctuations of the Raman intensities as the particles diffuse in and out of the small observation volume, and we analyse these temporal fluctuations via auto-correlation functions (the correlations of the Raman signal with itself, at points separate in time), see Fig. 1.

The correlation functions provide a characteristic time- value from which the nanoparticles’ diffusion coefficient

metal colloids are suspensions of nanometre-size gold or silver particles dispersed in a liquid, here water. Under illumination by light, these solutions present unique and amazing optical properties that depend strongly on the size, shape and state of aggregation of the particles. However in spite of intense research work dedicated to metal colloids, understanding their optical properties in relation with the geometry of the particles as well as investigating their aggregation processes still poses a highly challenging task. Combining correlation analysis with Raman scattering measurements answers this challenge.

metal colloids investigated by

enhanced- Raman correlation

spectroscopy

is inferred. Rotation of anisotropic clusters within the focus volume also provokes temporal fluctuations of the intensities, giving rise to a second characteristic time related to the rotational diffusion coefficient of the nanoparticles and/or clusters. From the values of the diffusion and rotation coefficients, and using hydrodynamic models, the size and shape of the nanoparticles/clusters can be identified.

As a proof-of-principle we studied a model system of aggregates of ~30 nm diameter silver nanoparticles where the coagulation of the nanoparticles was induced by p-mercaptobenzoic acid (MBA) molecules.

Depending on the MBA concentration, ranging from 1 to 10^-3 millimoles/litre, different aggregation stages could be achieved and identified. For the highest concentration in MBA, we showed that large, isotropic clusters with an effective hydrodynamic radius almost 30 times larger than that of the isolated nanoparticle were formed. At lower MBA concentrations the first stages of the aggregation process could be studied. Dimers and chains of four nanoparticles were observed. Finally, in the case of the lowest MBA concentration, no aggregation occurred and the diffusion of the isolated nanoparticles of the initial colloidal solution was observed.

These results show that Surface-Enhanced Raman Scattering correlation spectroscopy is a very promising and powerful technique for in-situ studies of this kind.

It enables the quantitative sizing and characterization of SERS-active aggregates and isolated nanoparticles.

It thus paves the way to the practical task of in-situ monitoring of nanoparticle aggregation processes, in conjunction with basic research into the optical properties of the as-formed clusters.

Fig. 1: At left: A microscope illuminates a colloidal solution of nanoparticles (grey circles) carrying adsorbed molecules (blue dots). The Surface Enhanced Raman Scattering by the molecules, in the backwards direction, is recorded as a function of time (top right). The auto-correlation function (bottom right) yields rotational and translational characteristic times related to the geometry of the diffusing objects.

CONTACT

Aude BARBARA aude.barbara@neel.cnrs.fr

FURTHER READING

“SERS correlation spectroscopy of silver aggregates in colloidal suspension: quantitative sizing down to a single nanoparticle”

a. Barbara, F. Dubois, a. Ibanez, L. Eng and P. Quémerais J. Phys. Chem. C 118, 17922 (2014).

“Non-resonant and non-enhanced raman correlation spectroscopy”

a. Barbara, F. Dubois, P. Quémerais and L. Eng Opt. Exp. 21, 15418 (2013).

CONTACT

Florence FERNANDEZ florence.fernandez@neel.cnrs.fr

FURTHER READING

neel.cnrs.fr/physiquarium

This dedicated visitors’ area aims to make the story of scientific advances- past and present- comprehensible and attractive, and to provide glimpses of future developments. It was planned as a gateway to the Institut NÉEL’s research work, adaptable for a varied public. It provides a pedagogical presentation of our work, including possibilities for simple demonstration experiments.

The

PHYSIQUARIUm, a visitors’ area

at the Institut NÉEL

A specific programme also exists with the École des Pupilles de l’Air de Grenoble in which their post- baccalaureate year students develop new pedagogical experiments, for presentation at the PHYSIQUARIUM and possible integration into its panoply of experiments.

The researchers responsible for the PHYSIQUARIUM are Julien Delahaye, Jean-Louis Hodeau and Pierre Molho. Many Institut NÉEL staff and Ph.D. students contribute regularly to its success, including by welcoming the visiting students into their laboratories.

Popularisation of science is an important mission and the PHYSIQUARIUM is a concrete example of the Institut NÉEL’s ongoing commitment to communicating with the public and especially to awakening young people to science. In return, our staff have been gratified by the response of teachers and their pupils. And for our Ph.D.

students, the PHYSIQUARIUM provides an opportunity to present – often for the first time – their discipline to a non-specialist audience.

In 2016, a fourth section will be added to the PHYSIQUARIUM, called “Microscopies” (to include optical, electron, and near-field microscopies), which is another major area of the Institut NÉEL’s research.

Opened in 2014, the PHYSIQUARIUM consists at present of three sections: Magnetism, Crystallography and Low Temperatures, three domains that are fundamental to much of the Institut NÉEL’s research.

Text, images and videos (accessed interactively via a touch-screen terminal), multiple posters and “hands- on” experiments introduce the visitor to the history of these three domains and present a selection of our recent discoveries and current work.

Various frameworks for visits are possible. In 2015, these have benefited, in particular: classes from regional high schools; the winners of the national “Olympiades de Physique” competition; the general public during Grenoble’s annual Fête de la Science; the undergraduate- level university students attending the International Nanosciences Summer School.

The first initiatives have interested some of the Institut NÉEL’s partners. Representatives from Okayama University have visited the PHYSIQUARIUM to prepare the visit of 15 Japanese high school students in 2016.

Locally, a formal collaboration has been started with our regional education authorities: a specific science-popularisation programme named “school@

physiquarium”. This programme is designed to actively engage high-school students, to stimulate their curiosity and sense of wonder, and to encourage them to apply all their powers of observation. Before the visits, the classes’ science teachers meet with Institut NÉEL personnel and a supervising teacher for a day of preparation at the PHYSIQUARIUM. This enables the teachers to plan the most appropriate experiments to undertake with their students.

Fig. 1: View of part of the visitor area.

Il est de notre responsabilité de vul- gariser nos travaux pour les rendre accessibles à un plus grand nombre.

«

»

Alain Schuhl, Director of the Institut NÉEL 2011-2014

Fig. 2: Observation.

force-field’s cartography (Fig. 1c), on each side of the optical axis.

Often in physics, the dimensionality of the problem strongly affects the phenomenology observed. This is the case in this experiment, since the abovementioned dynamical instability cannot occur in the “mono- dimensional” optomechanical systems studied previously (resonators forming one end of a Fabry-Perot cavity) where the forces exerted by optical fields are conservative.

The optomechanical investigation developed in this work represents a new tool for investigating confined force fields. More generally, performing vectorial force field cartography provides a new analytical tool for investigating the light-matter interaction.

In collaboration with colleagues at the Institut Lumière Matière (Lyon) we have developed an experiment to probe the vibrations of nanoscale mechanical oscillators by an optical technique. Specifically, our nano- mechanical oscillators are crystalline Silicon Carbide nanowires having ultralow masses, of one-tenth of a picogram (10^-16 kg). These oscillators are sensitive to forces in the attonewton range (10^-18 N). They are one million times more sensitive than the (already very sensitive) probes used in the Atomic Force Microscopes that have been revolutionizing the nanosciences in recent years. We have used this extreme sensitivity to measure and map the mechanical action of light.

A 30 µm long, 150 nm diameter nanowire (Fig. 1a) was immersed in a laser beam that was tightly focused, down to a 500 nm spot size (Fig. 1b). First, by measuring the fluctuations of the light scattered by the nanowire, we characterized the nanowire’s vibrations, whose fundamental eigenmodes have frequencies in the 10²- 10³ kHz range. We were able to measure the thermal vibrational noise of the nanoresonator, (its “Brownian motion”) over a very large dynamic range. This random motion determines the intrinsic force sensitivity of the nanowire, which amounts here to a few attonewtons at the temperature of the experiment, 300 K.

By moving the nanowire within the optical spot using a very precise piezo-electric positioning device, we then measured, in each point in space, the force exerted on the wire by the light field. Because the nanowire has such a tiny diameter, we could map the light-matter interaction with sub-wavelength spatial resolution (Fig. 1c). This represents a first step towards the investigation of vectorial force fields at the nanoscale, with unequalled sensitivity.

We also investigated how the presence of the optical force field can modify the nanowire’s dynamics. This perturbation presents a fundamental character since it represents the return action or “back-action” of the optical measurement process, to which all measurement systems are intrinsically submitted.

In particular, we have investigated how the detailed shape of the vectorial force field (its “topology”) was impacting the nanowire’s dynamics. We have discovered a novel dynamical instability of the nanowire, leading to very large self-sustained oscillations, that appears in regions of strong vorticity of the light field. The existence of these regions is a consequence of the non-conservative character of the light-matter interaction. The regions of large shear can be directly visualized in the measured

Because of their ultra-low mass, nanometre-scale objects can be used to measure applied forces with exceptional sensitivity, provided one can detect their vibrations. Here, a nanowire made of Silicon Carbide has been used to map the force field exerted by a focused laser beam. It is well known that the radiation pressure can have significant effects at the macroscopic and astronomical scales. What is radically new with nanowires is that we can study the mechanical action of light at a scale smaller than the optical wavelength. In addition, we can investigate the back action of the light on the nanowire’s vibrational dynamics.

Fig. 1: (a) Scanning Electron Microscope image of a 150 nm diameter Silicon Carbide nanowire, suspended at its top on a tungsten tip.

(b) The nanowire is piezo-positioned in a laser beam strongly focused by high numerical aperture microscope objectives. The nanowire’s vibrations modulate the transmitted light beam intensity, which provides a means to characterize the nanowire’s vibrations and measure its thermal noise.

(c) Due to a non-perfect cylindrical symmetry, each vibrational mode of the wire has two perpendicular components with slightly different frequencies.

By amplitude-modulating the laser beam in resonance with these frequencies, the z and x components of the force exerted by the laser on the wire can be determined and mapped with large signal to noise (to give the field of red arrows). At high laser power, a dynamical instability is observed in regions of large shear in the vector flow, the irregularities seen here on each side of the optical axis.

Probing light fields with nano-mechanical

oscillators

CONTACT

Olivier ARCIZET olivier.arcizet@neel.cnrs.fr Ph.D student:

arnaud GLOPPE

FURTHER READING

“Nano-optomechanics and topological backaction in a non-conservative radiation force field”

a. Gloppe, P. Verlot, E. Dupont-Ferrier, a. siria, P. Poncharal, G. Bachelier, P. Vincent, O. arcizet Nature Nanotechnology 9, 920 (2014).

Imaging inside Gallium Nitride wires for next generation

blue LEDs

magnetic domain walls are the transition regions between two magnetic domains with different magnetization directions. In ultra-thin films of a magnetic metal such as cobalt, having magnetization

perpendicular to the film plane, the domain walls can be very narrow: 5 - 20 nm. Such domain walls can be moved at speeds faster than 400 m/s by in-plane current pulses and could be used as carriers of binary information in future high-density magnetic storage devices. These high speeds are possible because the magnetic configuration within the domain walls has definite “chirality”: the magnetization rotates in the same direction in every wall in the cobalt layer.

Chiral domain walls

in thin magnetic films

for a rectangular Pt/Co/AlO_x microstructure. The magnetization of the cobalt is first saturated in the up direction. A magnetic field pulse is then applied in the down direction. With zero in-plane field (Fig. 2a), reversal of the cobalt’s magnetization starts at a defect in the centre of the film. However, in the presence of an in- plane field B_x , nucleation starts either at the left edge (Fig. 2b) or at the right edge (Fig. 2c) of the structure, depending on the direction of B_x .

The first observations of high-speed domain wall move- ment driven by an electrical current were made by Spin- tec Laboratory, Grenoble, in collaboration with the Institut NÉEL. This was in 0.6 nm (2 - 3 atomic layers) cobalt films sandwiched between a thicker layer of the non-magnetic metal platinum and an insulating layer of aluminium oxide (AlO_x ). Following that discovery, it was proposed (Thiaville et al. EPL 2012) that two effects induced by the Pt/Co inter face could explain such high propagation speeds.

First, the combination of a strong spin-orbit interaction in platinum and the breaking of the structural inversion symmetry by the interfaces can lead to a type of exchange interaction, called the Dzyaloshinskii-Moriya (DM) interaction, that favours a tilted alignment between neighbouring Co spins. This interaction can lead to domain walls where the magnetization rotates progressively in a plane perpendicular to the wall (see Fig. 1). This type of domain wall is well known (such walls are called Néel domain walls after Louis Néel) but in this system they have a special property: the sense of rotation is not random but is imposed by the sign of the DM exchange interaction, and is the same for all domain walls in the Co film. Hence the label “chiral domain walls”.

In this scenario, the second ingredient needed for efficient wall motion is a phenomenon called the Spin Hall effect that produces a spin-polarized electron current flowing from the Pt to the Co layer when an electrical current passes through the Pt layer. This spin current flow generates a strong torque that moves all the chiral walls in the same direction through the cobalt layer.

Fig. 1 illustrates the expected magnetization confi- guration of Néel walls with left-handed chirality. Notice in particular that the magnetization direction at the center of each domain wall is horizontal and opposite within the two, consecutive walls that separate down/up and up/down domains. We can make use of this property to prove that the domain walls are indeed chiral, by applying an in-plane magnetic field B_x : the energy of these walls will differ according to whether their mean magnetization is parallel or anti-parallel to B_x .

Fig. 2 shows images we have obtained using Kerr microscopy (magnetic microscopy with polarized light)

Fig. 2: Kerr-microscopy demonstration of the left-handed chiral nature of Néel domain walls in a Pt/Co(0.6nm)/AlO_x microstructure. A small, perpendicular, magnetic field pulse Bz is applied in the down direction to a layer that had been initially magnetized in the up direction (dark grey). In (a), with no in-plane field (Bx ), a small domain with down magnetization (light grey) nucleates at a defect.

With an in-plane field 260 mT directed right in (b), or left in (c), magnetization reversal starts along the left or right edge respectively, and spreads rapidly.

Fig. 1: Two Néel domain walls with fixed left-handed chirality in a thin Co layer having perpendicular magnetization along z. The magnetization rotates to the left (anticlockwise around y) in both walls. Thus, the mean magnetization has opposite directions along x in the two successive domain walls.

CONTACT

Stefania PIZZINI

stefania.pizzini@neel.cnrs.fr

FURTHER READING

“Chirality-induced asymmetric magnetic nucleation in Pt/Co/AlO_x ultrathin microstructures”

s. Pizzini, J. Vogel, s. rohart, L. Buda-Prejbeanu, E. Jué, O. Boulle, I.M. Miron, C.K. safeer, s. auffret, G. Gaudin, a. Thiaville Phys. Rev. Lett. 113, 047203 (2014).

These results confirm the presence of chiral Néel domain walls. By geometry, chiral walls created on the opposite edges have opposite in-plane magnetization (see Fig. 1).

The nucleation takes place at the side where the domain wall’s energy is lower, i.e. where its mean magnetization is parallel to B_x . A left-handed chirality (like in Fig. 1) can then be deduced from our measurements for the domains walls in the Pt/Co system.

The perpendicular field needed for nucleating an edge domain, i.e. to trigger reversal of the magnetization, depends strongly on the strength of B_x . By modelling this dependence we could obtain the strength of the Dzyaloshinskii-Moriya interaction in this system, which is D = 2.2 mJ/m². This is a high value and leads to a high stability of the domain walls in Pt/Co heterostructures, which is essential for high velocity propagation, making them a promising system for binary storage applications.

In the language of nano-technology, a “wire” is a very small crystalline object having length considerably greater than its diameter. Wire-based devices are a promising new route towards improved electronic and optoelectronic devices. Compared to present-generation nitride LEDs, which are conventional planar devices, wire devices have very desirable intrinsic properties such as small footprints and improved crystal quality. With careful design of donor and acceptor doping and alloy compositions, wires offer more versatility.

The building-block of optoelectronic applications is the p-n junction. In a LED, electrons from the n-type region and holes from the p-type region recombine at the junc- tion, emitting light. Especially promising are p-n junctions in wires grown with the “core-shell” geometry: An n-type core (doped with the donor impurity silicon) is totally enclosed by a p-type shell (doped with the acceptor im- purity magnesium). The shell is deposited on the lateral walls of the core and often also on the top of the core, see Fig. 1. Thus the p-n junction at the core-shell bound- ary has a three-dimensional (3D) character, very different

from that of planar diodes. The wire’s high ratio of surface to volume greatly increases the total area of the junction, and thus increases the proportion of active region in the device. This should alleviate the “efficiency droop” is- sue (a severe fall-off of efficiency at high current levels) in LEDs and will also be important for nitride-wire solar cells.

Semiconductors are both fascinating for physicists and extremely useful to society. The best-known semiconductor is Silicon, the material at the origin of the electronics and digital revolution of the second half of the 20^th century. At the present time, the nitride materials Gallium Nitride (GaN) and its alloys are beginning another semiconductor revolution due to the specific properties of their energy bandgaps that enable them to emit visible light with unparalleled efficiency. The inventions at the origin of the nitride blue LEDs (Light Emitting Diodes) that are now penetrating the mass market for lighting applications were recognized by the 2014 Nobel Prize for Physics. We focus here on a potential next generation of these nitride LEDs, based on dense arrays of nitride microwires.

Imaging inside Gallium Nitride wires for

next generation blue LEDs

Development of these nitride core-shell wire devices is hampered at present by a lack of precise knowledge about essential semiconductor properties along the bur- ied 3D p-n junction. New, specific tools are needed to investigate the physical properties of the wire at the na- nometric scale. We have now demonstrated how to im- age the electric field in a core-shell GaN wire and thus to determine the doping concentrations of the p-n junction.

Our diodes were wire arrays (Fig. 2) grown by the catalyst- free Metal-Organic Vapor-Phase Epitaxy technique on conducting GaN and Si substrates. The wires were grown and processed into device structures by colleagues at the CEA-LETI, Grenoble. To image the top and lateral junctions existing in the 3D diode structure, we exploited two powerful techniques available on a Scanning Electron Microscope, namely Electron-Beam Induced- Current (EBIC) and secondary-electron Voltage Contrast.

We cleaved the substrates to split some of the GaN wires down their length (see Figs 1 and 2). With this cross- sectional approach and applying the scanning electron- beam techniques, we can achieve a spatially resolved analysis of the entire p-n junction, with nanoscale resolution.

The EBIC images (Fig. 2) enabled us to measure, at each point in the 3D junction, the diffusion lengths of cur- rent carriers generated by the electron beam (typically

≈60 nm on the p-type side of the junction and ≈15 nm on the n-side). The measurements also provided the width of the carrier depletion zone (40−50 nm). Under reverse bias of the diode, Voltage-Contrast Imaging provided electrostatic maps of the local electric potentials, from which we could deduce acceptor and donor impurity doping levels in the vicinity of the 3D junction (typically N_a = 3 × 10¹⁸ cm⁻³ and N_d = 3.5 × 10¹⁸ cm⁻³ in both the top and the lateral junction).

These nanoscale probing techniques provide values of the key parameters – carrier diffusion lengths, widths of space charge regions, doping levels – essential for guiding the further development of core-shell wire devices including LEDs and photovoltaic solar cells.

Fig. 1: A cleaved, core-shell GaN wire, and its integration into a diode device for Electron-Beam Induced-Current (EBIC) measurements. The focussed electron beam creates electron-hole pairs or excitons in the GaN semiconductor; e-h pairs or excitons created in the p-n junction region are separated by the junction’s electric field, and a net electric current flows through the device.

Fig. 2: (a) Field Emission Scanning Electron Microscope (SEM) image at 10 keV for a processed, wire- based device, and the corresponding image (yellow colour) of the Electron Beam Induced Current intensity (scale bar: 10 μm). The wires in the front row were cleaved in half, so that the electron beam can probe inside them.

(b) A zoom onto the EBIC image shows lateral and top p-n junctions in a cleaved wire.

CONTACT

Fabrice DONATINI fabrice.donatini@neel.cnrs.fr Julien PERNOT

julien.pernot@neel.cnrs.fr Ph.D student:

Pierre TCHOuLFIaN

FURTHER READING

“Direct Imaging of p–n Junction in Core–Shell GaN Wires”

P. Tchoulfian, F. Donatini, F. Levy, a. Dussaigne, P. Ferret and J. Pernot Nano Lett. 14, 3491 (2014).

Over the last four decades, the size of a bit, the smallest logical unit in a computer, has decreased by more than two orders of magnitude and will soon reach a limit where quantum phenomena become important. Inspired by the power of quantum mechanics, researchers have already identified pure quantum systems that have controllable and readable discrete states, in analogy

with a classical bit. In this regard, the inherent spin of an atomic nucleus with its two (or more) eigenstates is a promising candidate. We have developed a molecular transistor which allows us to write quantum information onto a single nuclear spin purely by means of an electric field, as well as to do an electronic read-out of the quantum state. This novel approach opens a path to addressing and manipulating individual nuclear spins, that is to handle information localized in a very small space (a single atom), at high speed.

Coherent control of a single

nuclear spin with an electric field

atomic level. When we apply an electric field, it modifies the electronic wave functions of the terbium ion (a “Stark Effect”). This modification in turn changes the coupling of the electronic and nuclear spins, the “hyperfine interaction”. The altered hyperfine interaction is seen by the nucleus as a change of the effective magnetic field at the nucleus. Therefore an oscillating electric field is transformed into an oscillating effective magnetic field. If the frequency of this oscillation is brought into resonance with the spacing between the nuclear spin levels, a coherent manipulation of the nuclear spin states is possible. Furthermore, the hyperfine Stark effect also allows us to shift the resonance frequency of our nuclear quantum bit (“qubit”) by means of a static electric field, which can be easily generated by applying a back-gate voltage.

The use of oscillating electric fields to manipulate nuclear spins could facilitate the integration of multiple spin- qubits in a device architecture, since electric fields are much easier to focus and to shield than magnetic fields.

Moreover, the use of a static back-gate voltage to tune individual nuclear spin qubits in and out of resonance provides a possibility for addressing individual qubits using only a single-frequency microwave signal. Thus, single-molecule magnets are promising candidates for quantum information processing, and could lead to the appearance of a new field of research: molecular quantum electronics.

The magnetic moment of a nuclear spin is ten billion times smaller than the magnetic moment of one bit of a modern hard drive. To succeed in detecting such a tiny signal, we first developed a very sensitive magnetic field sensor – the Single-Molecule Magnet (SMM) spin- transistor (Fig. 1). The heart of this device is a single metalorganic molecule. It is a terbium “double-decker”

single-molecule magnet, which works as amplifier and detector for the terbium ion’s nuclear spin.

To amplify the nuclear spin signal, we make use of the

“hyperfine” interaction between the terbium’s nuclear spin and its electronic spin. By mapping the nuclear spin state onto the terbium’s electronic spin state we amplify the magnetic signal by a factor of 10000. In a second stage of amplification we read-out the nuclear spin dependent magnetic moment of the terbium ion via the tunnel current through the molecular transistor. This is possible as the terbium’s electronic spin interacts strongly with the tunnelling electrons via its exchange coupling with those electrons.

In recent experiments at very low temperature (30 mK), we were able to show that the terbium’s nuclear spin can be manipulated purely by an oscillating electric field (the microwave field represented in Fig. 1). This is remarkable because the magnetic dipole associated with the nuclear spin is, to first order, by its very nature, sensitive to magnetic fields only.

Our experiment is explained by the hyperfine Stark effect, which acts as a magnetic field transducer at the

Fig. 1: The single-molecule magnet spin-transistor. The heart of the device is the terbium “double-decker“

single-molecule magnet: two planar organic molecules enclosing a Terbium ion. The “magnet“ is connected to two metallic leads (gold atoms) deposited on a substrate. We apply a voltage across the two contacts to establish a tunnelling current through the molecule. The read-out of the (I=3/2) ¹⁵⁹Tb nucleus’s four spin states (depicted in the enlargement as coloured rings) makes use of a change in this tunnelling current that occurs at well defined values of an applied external magnetic field. At these field values, the nuclear spin state is resonant with the microwave electric field pulse (blue wave).

CONTACT

Franck BALESTRO franck.balestro@neel.cnrs.fr Wolfgang WERNSDORFER wolfgang.wernsdorfer@neel.cnrs.fr Ph.D student:

Stefan THIELE

FURTHER READING

“Electrically driven nuclear spin resonance in single-molecule magnets”

S. Thiele, F. Balestro, r. Ballou, s. Klyatskaya, M. ruben, W. Wernsdorfer Science 344, 1135 (2014).